The possibility of a worldwide influenza

Transcription

1 Pandemic Economics: The 1918 Influenza and Its Modern-Day Implications Thomas A. Garrett Many predictions of the economic and social costs of a modern-day pandemic are based on the effects of the influenza pandemic of Despite killing 675,000 people in the United States and 40 million worldwide, the influenza of 1918 has been nearly forgotten. The purpose of this paper is to provide an overview of the influenza pandemic of 1918 in the United States, its economic effects, and its implications for a modern-day pandemic. The paper provides a brief historical background as well as detailed influenza mortality statistics for cities and states, including those in the Eighth Federal Reserve District, that account for differences in race, income, and place of residence. Information is obtained from two sources: (i) newspaper articles published during the pandemic and (ii) a survey of economic research on the subject. (JEL I1, N0, R0) Federal Reserve Bank of St. Louis Review, March/April 2008, 90(2), pp The possibility of a worldwide influenza pandemic in the near future is of growing concern for many countries around the globe. The World Bank estimates that a global influenza pandemic would cost the world economy $800 billion and kill tens of millions of people (Brahmbhatt, 2005). Researchers at the U.S. Centers for Disease Control and Prevention (CDC) calculate that deaths in the United States could reach 207,000 and the initial cost to the economy could approach $166 billion, or roughly 1.5 percent of GDP (Meltzer, Cox, and Fukuda, 1999). The U.S. Department of Health and Human Services paints a more dire picture up to 1.9 million dead in the United States and initial economic costs near $200 billion (U.S. Department of Health and Human Services, 2005). The long-run costs of a modern-day influenza pandemic are expected to be much greater. Although researchers and public officials can only speculate on the likelihood of a global influenza pandemic, many of the worst-case scenario predictions for a current pandemic are based on the global influenza pandemic of That pandemic killed 675,000 people in the United States (nearly 0.8 percent of the 1910 population), a greater number than U.S. troop deaths in World War I (116,516) and World War II (405,399) combined. 1 Roughly 40 million people died worldwide from the early spring of 1918 through the late spring of In all of recorded history, only the Black Death that occurred throughout Europe from 1348 to 1351 is estimated to have killed more people (roughly 60 million) over a similar time period (Bloom and Mahal, 1997). The years 1918 and 1919 were difficult not only because of the influenza pandemic, but because these years also marked the height of 1 See Potter (2001) for a discussion of 1918 influenza pandemic mortalities. U.S. troop mortality data can be found at 2 Although 40 million is the commonly accepted number of worldwide deaths from the pandemic, it is likely an underestimate given the lack of adequate recordkeeping in many parts of the world. Thomas A. Garrett is an assistant vice president and economist at the Federal Reserve Bank of St. Louis. Lesli S. Ott provided research assistance. 2008, The Federal Reserve Bank of St. Louis. Articles may be reprinted, reproduced, published, distributed, displayed, and transmitted in their entirety if copyright notice, author name(s), and full citation are included. Abstracts, synopses, and other derivative works may be made only with prior written permission of the Federal Reserve Bank of St. Louis. FEDERAL RESERVE BANK OF ST. LOUIS REVIEW MARCH/APRIL

2 U.S. involvement in World War I. Given the magnitude and the concurrence of both the influenza pandemic and World War I, one would expect volumes of research on the economic effects of each event. Although significant literature on the economic consequences of World War I does exist (Rockoff, 2004), the scope of research on the economic effects of the 1918 influenza pandemic is scant at best. Most research has focused on the health and economic outcomes of descendents of pandemic survivors and the mortality differences across socioeconomic classes. (See, for example, Keyfitz and Flieger, 1968; Noymer and Garenne, 2000; Almond, 2006; and Mamelund, 2006.) Certainly an event that caused 40 million worldwide deaths in a year should be closely examined not only for its historical significance, but also for what we can learn (in the unfortunate chance the world experiences another influenza pandemic). This paper discusses some of the economic effects of the 1918 influenza pandemic in the United States. The first section discusses demographic differences in pandemic mortalities: Were deaths higher in cities than in rural areas? Did deaths differ by race? Detailed influenza mortality data at various geographic and demographic levels at the time of the pandemic are available. The presentation of mortality data series allows for an almost unlimited number of comparisons and analyses that afford the reader the opportunity to study the available data and generate his own analyses and conclusions in addition to those presented here. Evidence on the effects of the pandemic on business and industry is obtained from newspaper articles printed during the pandemic, with most of the articles appearing in newspapers from the Eighth Federal Reserve District cities of Little Rock, Arkansas, and Memphis, Tennessee. Newspaper articles from the fall of 1918 were used because of the almost complete absence of economic data from the era, such as data on income, employment, sales, and wages. This absence of data, especially at local levels (e.g., city and county), is a likely reason for the scarcity of economic research on the subject, although several studies that have used available economic data are reviewed here. Although the influenza pandemic occurred nearly 90 years ago in a world that was much different from today s, the limited economic data and more readily available mortality data from the time of the event can be used to make reasonable inferences about the economic and social consequences of a modern-day pandemic. Despite technological advances in medicine and greater health coverage throughout the twentieth century, deaths from a modern-day pandemic are also likely to be related to race, income, and place of residence. Thus, the geographic and demographic differences in pandemic mortalities from 1918 can shed light on the possible effects of a modernday pandemic, a point that is taken up in the final section of the paper. OVERVIEW OF THE 1918 INFLUENZA PANDEMIC The influenza pandemic in the United States occurred in three waves during 1918 and The first wave began in March 1918 and lasted throughout the summer of The more devastating second and third waves (the second being the worst) occurred in the fall of 1918 and the spring of 1919 as the pandemic spread across the country: Spanish influenza moved across the United States in the same way as the pioneers had, for it followed their trails which had become railroads the pandemic started along the axis from Massachusetts to Virginia leaped the Appalachians positioned along the inland waterways it jumped clear across the plains and the Rockies to Los Angeles, San Francisco, and Seattle. Then, with secure bases on both coasts...took its time to seep into every niche and corner of America. (Crosby, 2003, pp ) But the pandemic s impact on communities and regions was not uniform across the country. For example, Pennsylvania, Maryland, and 3 For much more information on the influenza pandemic, including its origins, see Crosby (2003) and Barry (2004). 76 MARCH/APRIL 2008 FEDERAL RESERVE BANK OF ST. LOUIS REVIEW

3 Colorado had the highest mortality rates, but these states had very little in common. Arguments have been made that mortality rates were lower in later-hit cities because officials in these cities were able to take precautions to minimize the impending influenza, such as closing schools and churches and limiting commerce. The virulence of the influenza, like a typical influenza, weakens over time, so the influenza that struck the East Coast became somewhat weaker by the time it struck the West Coast. But these reasons cannot completely explain why some cities and regions experienced high mortality rates while others were barely hit with the influenza. 4 The global magnitude and spread of the pandemic was exacerbated by World War I, which itself is estimated to have killed roughly 10 million civilians and 9 million troops. Not only did the mass movement of troops from around the world lead to the spread of the disease, tens of thousands of Allied and Central Power troops died as a result of the influenza pandemic rather than combat (Ayres, 1919). Although combat deaths in World War I did increase the mortality rates for participating countries, civilian mortality rates from the influenza pandemic of 1918 were typically much higher. For the United States, estimates of combat-related troop mortalities are about one-tenth that of civilian mortalities from the 1918 influenza pandemic. Mortality rates from a typical influenza tend to be the greatest for the very young and the very old. What made the 1918 influenza unique was that mortality rates were the highest for the segment of the population aged 18 to 40, and more so for males than females of this age group. In general, death was not caused by the influenza virus itself, but by the body s immunological reaction to the virus: Individuals with the strongest immune systems were more likely to die than individuals with weaker immune systems. 5 One 4 Much research has been conducted over the past decades to provide insights into why the pandemic had such different effects on different regions of the country (see, for example, Crosby, 2003, and Barry, 2004). One commonly held reason is the response of local governments to the influenza in their communities, e.g., partial versus full quarantines. 5 The lungs typically filled with fluid and the victim drowned or died of pneumonia. See Barry (2004). source reports that, of 272,500 male influenza deaths in 1918, nearly 49 percent were aged 20 to 39, whereas only 18 percent were under age 5 and 13 percent were over age The fact that males aged 18 to 40 were the hardest hit by the influenza had serious economic consequences for the families that had lost their primary breadwinner. Despite the severity of the pandemic, it is reasonable to say that the influenza of 1918 has almost been forgotten as a tragic event in American history. This is not good, as learning from past pandemics may be the only way to reasonably prepare for any future pandemics. Several factors may explain why the influenza pandemic of 1918 has not received a notable place in U.S. history. 7 First, the pandemic occurred at the same time as World War I. The influenza struck soldiers especially hard, given their living conditions and close contact with highly mobile units. Much of the news from the day focused on wartime events overseas and the current status of American troops. Thus, the pandemic and World War I were seen almost as one event rather than two separate events. Second, diseases of the day such as polio, smallpox, and syphilis were incurable and a permanent part of society. Influenza, on the other hand, swept into communities, killed members of the population, and was gone. Finally, unlike polio and smallpox, no famous people of the era died from the influenza; thus, there was no public perception that even the politically powerful and rich and famous were susceptible to the virus. Despite its lack of historical prominence, the influenza pandemic of 1918 created significant economic and social effects, even if these were short-lived. In select areas, increasing body counts overwhelmed city and medical officials. In some cities, such as Philadelphia, bodies lay along the streets and in morgues for days, similar to medieval Europe during the Black Death. In light of the potential economic turmoil and human suffering, an understanding of the state and fed- 6 The 272,500 deaths are from a sample of about 30 states. See Crosby (2003, p. 209). 7 See Crosby (2003, pp ). FEDERAL RESERVE BANK OF ST. LOUIS REVIEW MARCH/APRIL

4 eral government response to the 1918 pandemic may also shed some light into what government at any level can do, if anything, to prevent or minimize a modern-day pandemic. PANDEMIC MORTALITIES IN THE UNITED STATES Data on mortalities from the 1918 influenza pandemic are found in Mortality Statistics, an annual publication that is released by the U.S. Census Bureau. 8 Mortalities resulting from hundreds of causes of death are listed (depending on the level of data aggregation) and are also broken down, in some cases, by age, race, and sex. Data are available at the national, state, and municipal levels and may be available by week, month, and year. In terms of coverage, (a)ll death rates are based on total deaths, including deaths of nonresidents, deaths in hospitals and institutions, and deaths of soldiers, sailors, and marines (U.S. Department of Commerce, Bureau of the Census, 1922, p. 9). 9 The mortality rates used in this study represent deaths from both influenza and pneumonia in a given year because it is not believed to be best to study separately influenza and the various forms of pneumonia for doubtless many cases were returned as influenza when the deaths were caused by pneumonia and vice versa (U.S. Department of Commerce, Bureau of the Census, 1921, p. 28). 10 Although Mortality Statistics provides a remarkable number of statistics, a major disadvantage of the earlier reports is that, in the 1910s, data coverage is for only 75 to 80 percent of the total population. This is because the U.S. Census Bureau acquired the mortality data over time from a registration area that consisted of a growing group of states. So mortality data for certain states are not consistently available over time. For the 8 Copies of the historical reports are available at the CDC, National Center for Health Statistics, or at pubs/pubd/vsus/historical/historical.htm. Mortalities are likely to be underestimated, as overburdened health professionals stopped recording deaths during the peak of the pandemic. 9 Hereafter, this reference will be cited as Mortality Statistics Hereafter, this reference will be cited as Mortality Statistics purposes of this article, influenza mortality data for the 1910s are available for about 30 states and encompass, on average, about 79.5 percent of the U.S. population. A casual look at the states that did and did not report mortality information does not reveal any systematic differences across each group of states with regard to population, income, and race. So the available mortality statistics are unlikely to provide a biased picture of influenza mortalities. The following sections report select influenza mortality data at various levels of data aggregation (city and state), by race (white and non-white), and by residence (urban versus rural). The abundance of mortality statistics makes it impossible to use all existing data in a single report. However, the statistics used here do reveal some general mortality patterns that provide insights into which groups of people may be most/least affected by a modern-day pandemic, as well as how influenza mortalities differed across cities and states. State and City Pandemic Mortalities Pandemic mortality rates (per 100,000) for 27 states are shown in Table 1 for 1918 and The mortality rate for 1915 is also included and the ratio of 1918 mortalities to 1915 mortalities is calculated to reveal the deaths in 1918 relative to a non-pandemic year. 11 For the states shown in Table 1, Pennsylvania, Maryland, and New Jersey had the highest mortality rates in 1918, whereas Michigan, Minnesota, and Wisconsin had the lowest. The pandemic also lasted throughout the spring of 1919, so the ranking of states in 1918 does not reflect total mortalities in each state for the entire pandemic (although the rankings do remain similar). The ratio of the 1918 mortality rate to the 1915 mortality rate ranges from a low of 3.2 (Indiana and New York) to a high of 6.5 (Montana). One caveat is that an equal increase in mortalities for a lower-population state and a higher-population state will result in a greater mortality ratio for the 11 The non-pandemic year is assumed to be a normal influenza year. Later analyses of city influenza mortality rates use actual data on normal and excess mortality rates rather than assuming all years except 1918 and 1919 were normal. 78 MARCH/APRIL 2008 FEDERAL RESERVE BANK OF ST. LOUIS REVIEW

5 Table 1 Influenza Mortality Rates (per 100,000) for Select States Ratio of 1910 Area Population Mortality Mortality Mortality 1918 and 1918 State Population (miles 2 ) density rate rate rate 1915 rates Rank California 2,377, , Colorado 799, , Connecticut 1,114,756 4, Indiana 2,700,876 36, Kansas 1,690,949 81, Kentucky 2,289,905 40, Maine 742,371 29, Maryland 1,295,346 9, Massachusetts 3,366,416 8, Michigan 2,810,173 57, Minnesota 2,075,708 80, Missouri 3,293,335 68, Montana 376, , New Hampshire 430,572 9, New Jersey 2,537,167 7, New York 9,113,614 47, North Carolina 2,206,287 48, Ohio 4,767,121 40, Pennsylvania 7,665,111 44, Rhode Island 542,610 1, South Carolina 1,515,400 30, * * 10 Tennessee 2,184,789 41, * * 21 Utah 373,351 82, Vermont 355,956 9, Virginia 2,061,612 40, Washington 1,141,990 66, Wisconsin 2,333,860 55, NOTE: Mortality rates are from Mortality Statistics 1920 (U.S. Department of Commerce, Bureau of the Census, 1922) and include mortalities from influenza and pneumonia. *Mortalities for South Carolina and Tennessee in 1915 are 1916 and 1917 figures, respectively. Population density is population per square mile. FEDERAL RESERVE BANK OF ST. LOUIS REVIEW MARCH/APRIL

6 Table 2 Correlations of State Characteristics with Influenza Mortalities 1915 Mortality rate 1918 Mortality rate Ratio of 1918 and 1915 rates Density (population/miles 2 ) 0.632* 0.447* Area (miles 2 ) 0.566* Population NOTE: *Denotes statistical significance at 5 percent level or better. Correlations are based on the data in Table 1 (n = 27). lower-population state because the increase in mortalities is a greater percentage of its population. Nevertheless, a comparison of 1915 mortality rates with those in 1918 and 1919 clearly reveals how much more severe the 1918 influenza was relative to influenza in a non-pandemic year. Evidence suggests that influenza mortality rates had no relationship with state economic conditions, climate, or geography (see Crosby, 2003, and Brainerd and Siegler, 2003). After providing a survey of anecdotal evidence and conducting statistical analyses, Brainerd and Siegler (2003, p. 7) conclude that the statistical evidence also supports the notion of influenza mortality as an exogenous shock to the population. However, because influenza is spread by close human contact, influenza infection and mortality rates are commonly greater in more densely populated areas. It thus serves as an interesting exercise to see whether there is a relationship between pandemic mortalities and state population size and population density. It is also worth exploring whether the relationships are different in a pandemic year compared with a non-pandemic year. Table 2 thus presents pairwise correlations (and their statistical significance) between state population, area, and population density and 1915 mortality rates, 1918 mortality rates, and the ratio of the two mortality rates. The correlations shown in Table 2 reveal that mortality rates in 1915 were greater in more densely populated states (0.632), but lower in larger states ( 0.566). State size had no significant correlation with 1918 mortality rates, but population density was correlated with 1918 mortality rates (0.447). Note, however, that the correlation between mortality rates and density is less for 1918 mortalities than for 1915 mortalities. This finding, in addition to the fewer significant correlations (albeit just one fewer), suggest that state size and population density had less influence on mortality rates in 1918 than in Thus, as suggested by earlier research, the location of individuals was less of a factor in dying from the 1918 influenza than from a non-pandemic influenza. 12 Furthermore, the ratio of mortality rates had no relationship with state size, population, or population density, as seen in the last column of Table 2. Mortality statistics for 49 cities are listed in Table 3. As seen in the state-level statistics, influenza mortalities in U.S. cities during the pandemic were three to five times higher, on average, than during a non-pandemic year (1915). There is slightly more variation in 1918 mortality rates across cities (σ = 182) than across states (σ = 146). The cities with the highest 1918 mortality rates (Pittsburgh, Scranton, and Philadelphia) are all located in Pennsylvania, and the cities with the lowest rates (Grand Rapids, Minneapolis, and Toledo) are all located in the Midwest. It is possible to get an idea of the influenza s effect on rural areas versus urban areas by calculating the average 1918 mortality in all cities in a state (for which mortality data were available) and then dividing by the state-level mortality rate See Crosby (2003). 13 Mortality rates for 64 cities (49 of which appear in Table 3) were used in the calculations. The other 15 cities were not included in Table 3 because of missing data. The mortality rates for these 15 cities can be obtained from the author. 80 MARCH/APRIL 2008 FEDERAL RESERVE BANK OF ST. LOUIS REVIEW

7 Table 3 Influenza Mortality Rates (per 100,000) for Select Cities Ratio of City Population Mortality rate Mortality rate Mortality rate and 1915 rates Rank Albany, New York 100, Atlanta, Georgia 154, Baltimore, Maryland 558, Birmingham, Alabama 132, Boston, Massachusetts 670, Bridgeport, Connecticut 102, Buffalo, New York 423, Cambridge, Massachusetts 104, Chicago, Illinois 2,185, Cincinnati, Ohio 353, Cleveland, Ohio 560, Columbus, Ohio 181, Dayton, Ohio 116, Denver, Colorado 213, Detroit, Michigan 465, Fall River, Massachusetts 119, Grand Rapids, Michigan 112, Indianapolis, Indiana 233, Jersey City, New Jersey 267, Kansas City, Missouri 248, Los Angeles, California 319, Lowell, Massachusetts 106, Memphis, Tennessee 131, Milwaukee, Wisconsin 373, Minneapolis, Minnesota 301, Nashville, Tennessee 110, New Haven, Connecticut 133, New Orleans, Louisiana 339, New York, New York 4,766, Newark, New Jersey 347, Oakland, California 150, Omaha, Nebraska 124, Paterson, New Jersey 125, Philadelphia, Pennsylvania 1,549, Pittsburgh, Pennsylvania 533, , Portland, Oregon 207, Providence, Rhode Island 224, Richmond, Virginia 127, Rochester, New York 218, San Francisco, California 416, Scranton, Pennsylvania 129, Seattle, Washington 237, Spokane, Washington 104, St. Louis, Missouri 687, St. Paul, Minnesota 214, Syracuse, New York 137, Toledo, Ohio 168, Washington, D.C. 331, Worcester, Massachusetts 145, NOTE: Mortality rates are from Mortality Statistics 1920 and include mortalities from influenza and pneumonia. FEDERAL RESERVE BANK OF ST. LOUIS REVIEW MARCH/APRIL

8 Table 4 City Influenza Mortality Rate Relative to State Mortality Rate (1918) State These ratios are shown in Table 4. A ratio greater than 1 suggests influenza deaths were, on average, greater in a state s cities than in the rural areas of the state and vice versa for a ratio less than 1. As seen in Table 4, most of the ratios are greater than 1, with some much greater than 1 (Missouri, Kansas, and Tennessee), thus revealing that cities in their respective state had higher mortality rates than rural areas of that state. This finding supports the positive correlation between population density and influenza mortalities shown in Table 2. Influenza Mortalities and Race Average of cities relative to state Michigan 0.89 Colorado 0.95 California 1.01 New York 1.02 Maryland 1.04 Massachusetts 1.06 Connecticut 1.07 Washington 1.11 Pennsylvania 1.11 Minnesota 1.11 Indiana 1.13 New Jersey 1.16 Wisconsin 1.17 Virginia 1.17 Ohio 1.19 Missouri 1.32 Kansas 1.58 Tennessee 1.66 Influenza mortalities by race are available for some cities in the United States, although the racial breakdown is not as detailed as it is for modern-day mortality statistics. Mortality statistics for 1918 are provided on the basis of white and non-white. Table 5 presents a breakdown of white and non-white mortality rates (per 100,000 for each racial group) for 14 U.S. cities. For each racial group, influenza mortality rates for 1915 are also included so a comparison can be made between a pandemic year and a non-pandemic year. The first six columns of Table 5 clearly show that non-white influenza mortalities are higher than white influenza mortalities in both pandemic and non-pandemic years (except for Kansas City in 1918). Whites experienced relatively higher mortality during the pandemic year 1918 (compared with the non-pandemic year 1915) than did non-whites. It is likely that racial differences in influenza mortality rates reflect, to some degree, differences in population density (as seen in Table 2) and geography (as seen in Table 4). Data on white and non-white populations as well as rural and urban residences for several decennial Census years are shown in Table 6. In 1910, the great majority of the urban population (having a higher population density than rural areas) in the United States was white (over 90 percent). This offers some explanation as to why whites as a group had a much larger increase in influenza mortalities during the pandemic than did non-whites. But, the decline in the strength of the mortality/density relationship in 1918 compared with that of 1915 (see Table 2) suggests that urban location alone cannot account for the relatively large increase in influenza mortalities among whites. What does this imply if an influenza pandemic struck today? The last two columns of Table 6 reveal that the non-white population in the United States has become much more urban (27 percent in 1910 and 91 percent in 2000) compared with the white population (49 percent in 1910 and 75 percent in 2000). However, the fact that both racial groups are becoming more urban does not bode well for either group because population density will certainly be a significant determinant of mortality. However, a modern-day pandemic may result in greater non-white mortality rates because a greater percentage of the non-white population in the United States lives in urban areas. 82 MARCH/APRIL 2008 FEDERAL RESERVE BANK OF ST. LOUIS REVIEW

9 Table 5 Influenza Mortality Rate By Race and City, 1915 and 1918 White, Non-white White Non-white White, White Non-white White, 1915, 1915, mortality mortality as percent mortality mortality as percent as percent as percent rate rate of non-white rate rate of non-white of white of non-white City Birmingham , Atlanta Indianapolis Kansas City, Missouri Louisville 1, , New Orleans , Baltimore , Memphis Nashville , Dallas * 149.8* 45.3* 11.9* 17.7* Houston * 143.9* 68.1* 20.2* 23.3* Norfolk Richmond Washington, D.C NOTE: *Mortality rates for Dallas and Houston for 1915 are 1916 and 1917 figures, respectively. Table 6 Location and Race, White Non-white Percent of Percent of as percent of as percent of white population non-white population Year U.S. urban population U.S. urban population that is urban that is urban SOURCE: Population data are from Historical Statistics of the United States, U.S. Census. FEDERAL RESERVE BANK OF ST. LOUIS REVIEW MARCH/APRIL